Nucleotide sequence analysis using correlation images

An application area for computer visualization, namely, nucleotide sequence analysis, and a visualization technique, termed a correlation image (CI), are presented. A CI is produced by comparing two sequences in all possible alignments, with each image row corresponding to a distinct alignment. Information of interest to molecular biologists (e.g., inverted and direct repeats, insertions, deletions, homology divergence) are clearly manifested as highly distinct geometric patterns in the image. A typical CI may represent several megabytes of information, yet the display can easily be created and scanned in seconds. The derivation and interpretation of CIs are shown. They are compared with other nucleotide sequence analysis techniques. The power of CIs in uncovering homologies is illustrated.<
>